September 1, 2015

An optimisation consultancy

 

The Situation

Our client is a London based food delivery company who struggled to meeting delivery times as their number of customers grew.

The Client’s Business Goals

  • To use historic data to divide London into delivery zones
    • To have a courier level delivery plan, with delivery sequence, within 2 hours of a morning order
    • The delivery sequence needed to integrate within their existing logistics system

Our Solution

  • In our initial research we found that the this scenario was comparable to the travelling sales person problem, a problem with no know solution
    • The only way to solve the problem completely was to test all delivery sequences, but it is not possible to test all combinations within the time limit with current technology
    • Our next step was to define the parameters of the optimisation, deciding which rules were fixed (e.g. time limit), and which could change (number of couriers)
    • Using our knowledge of decision maths algorithms such as minimum spanning tree, mathematical modelling, and process optimisation we customised our own solution to fit within the parameters
    • After rigorous testing we found that we could achieve at least 95% of the optimal solution within the time limits

Key benefits

  • The solution minimised the amount of resource required, whilst maintaining deliver SLA’s
    • A customised solution integrated within their existing infrastructure, creating a seamless tool
    • Our client was provided feeds of the data so that delivery orders were known upfront, which helped packing the delivery vehicle
    • By taking a pragmatic approach we avoided costly R&D costs, and could deliver in a timely manner